Drug Class Review on Proton Pump Inhibitors

Transcription

1 Drug Class Review on Proton Pump Inhibitors Evidence Tables July 2006 Original Report Date: November 2002 Update 1 Report Date: April 2003 Update 2 Report Date: April 2004 Update 3 Report Date: May 2005 A literature scan of this topic is done periodically The purpose of this report is to make available information regarding the comparative effectiveness and safety profiles of different drugs within pharmaceutical classes. Reports are not usage guidelines, nor should they be read as an endorsement of, or recommendation for, any particular drug, use or approach. Oregon Health & Science University does not recommend or endorse any guideline or recommendation developed by users of these reports. Marian S. McDonagh, PharmD Susan Carson, MPH Oregon Evidence-based Practice Center Oregon Health & Science University Mark Helfand, MD, MPH, Director Copyright 2006 by Oregon Health & Science University Portland, Oregon All rights reserved.

2 Note: A scan of the medical literature relating to the topic is done periodically (see for scanning process description). The governance group elected to proceed with another update of this report. Please see timeline on the DERP website for details on the date of its release. Prior versions of the report can be accessed at the DERP website.

3 TABLE OF CONTENTS Evidence Tables Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI...3 Evidence Table 2. Quality assessment of included trials...63 Evidence Table 3. Non-erosive GERD short-term trials...71 Evidence Table 4. Erosive GERD relapse prevention...81 Evidence Table 5. Non-erosive GERD relapse prevention...91 Evidence Table 6. Randomized controlled trials of esophagitis treatment in children...93 Evidence Table 7. Randomized controlled trials of duodenal ulcer treatment: PPI vs PPI...97 Evidence Table 8. Duodenal ulcer recurrence rates on maintenance therapy Evidence Table 9. Randomized controlled trials of gastric ulcer treatment Evidence Table 10. Randomized controlled trials of NSAID-induced ulcer treatment Evidence Table 11. Randomized controlled trials of PPIs for prevention of NSAID-induced ulcer Evidence Table 12. Adverse effects in short-term RCTs: PPI vs PPI Funding: The funding source, the Center for Evidence-based Policy, is supported by 17 organizations, including 15 state Medicaid programs. These organizations selected the topic and had input into the Key Questions for this review. The content and conclusions of the review are entirely determined by the Evidence-based Practice Center researchers. The authors of this report have no financial interest in any company that makes or distributes the products reviewed in this report Suggested citation for this report: McDonagh MS, Carson S. Drug Class Review on Proton Pump Inhibitors Proton Pump Inhibitors Page 2 of 168

4 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Adachi et al, 2003 Population, Setting 85 patients at 6 medical institutions in Japan. Mean age 66 (SD 13); 51% male; 100% Asian Number Screened, Eligible, Enrolled, Esophagitis Grade (Grading Withdrawn, Lost to Criteria), Other Characteristics Followup Healing Rate at 4 Weeks Healing Rate at 8 Weeks Grade A: 24% Not reported (Per protocol analysis on 76 Grade B: 53% patients): Grade C: 21% omeprazole 20 mg: 85.7% Grade D: 2% lansoprazole 30 mg: 85% (Los Angeles classification) rabeprazole 20 mg: 92.9% 42% h. Pylori positive (NS) Screened NR/eligible NR/85 enrolled 20% of lansoprazole group lost to f/u for endoscopy vs 7% in other groups; but no loss to f/u for reporting of symptoms 85 analyzed for symptoms, 76 for endoscopy Proton Pump Inhibitors Page 3 of 168

5 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Adachi et al, 2003 Symptoms at 4 Weeks Symptoms at 8 Weeks Results by Baseline Severity Withdrawals Due to Adverse Events (Results reported graphically only) Heartburn score significantly lower in rabeprazole group after 2 days than lansoprazole or omeprazole (p=0.045). Differences disappeared by day 5. No significant differences in acid reflux scores. Not reported Not reported Not reported Proton Pump Inhibitors Page 4 of 168

6 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Adachi et al, 2003 Quality rating Funding source Fair: Ministry of open-label, loss to f/u higher in lansoprazole Education, Science, group for healing (20% vs 7%), but okay for and Culture of symptoms; randomization method not reported Japan Proton Pump Inhibitors Page 5 of 168

7 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Bardhan et al, 2001 Population, Setting 328 patients at 23 centers in Great Britain, the Republic of Ireland, and South Africa. Mean age 44.6 (SD 13.3) in pantoprazole group, 45.2 (SD14.4) in omeprazole group. 52.4% of pantoprazole, 64% of omeprazole group males. Race/ethnicity not reported. Number Screened, Eligible, Enrolled, Esophagitis Grade (Grading Withdrawn, Lost to Criteria), Other Characteristics Followup Healing Rate at 4 Weeks Healing Rate at 8 Weeks 100% Grade I (Savary-Miller classification) Screened NR/eligible NR/328 enrolled/ 327 analyzed Intention-to-treat (N=327): pantoprazole 20 mg: 77% omeprazole 20 mg: 81% Per-protocol (N=264): pantoprazole 20 mg: 84% omeprazole 20 mg: 89% Intention-to-treat (N=327): pantoprazole 20 mg: 81% omeprazole 20 mg: 88% (NS) Per-protocol (N=264): pantoprazole 20 mg: 90% omeprazole 20 mg: 95% (NS) Proton Pump Inhibitors Page 6 of 168

8 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Bardhan et al, 2001 Symptoms at 4 Weeks Symptoms at 8 Weeks Results by Baseline Severity pantoprazole 20 mg vs omeprazole Not reported Relief of acid eructation, heartburn and pain on 20 mg swallowing was similar in the two treatment Symptom relief (all main symptoms) groups at 2 and 4 weeks, irrespective of severity 2 weeks: 70% vs 79% at baseline. 4 weeks: 77% vs 84% A higher proportion with mild symptoms at entry Acid eructation had relief compared with patients with severe 2 weeks: 79% vs 88% symptoms, and this was similar for both 4 weeks: 84% vs 87% treatments. Heartburn 2 weeks: 79% vs 86% 4 weeks: 83% vs 87% Pain on swallowing 2 weeks: 83% vs 87% 4 weeks: 87% vs 97% (All NS) Withdrawals Due to Adverse Events Not reported Proton Pump Inhibitors Page 7 of 168

9 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Bardhan et al, 2001 Quality rating Funding source Fair-Poor: Byk Gulden open-label, randomization, allocation (Germany) concealment method not reported, more pharmaceutical smokers in pantoprazole group (31% vs 22%), more males in omeprazole group (64% vs 52%) Proton Pump Inhibitors Page 8 of 168

10 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Chen et al, 2005 Population, Setting 48 patients at a single center in Taiwan. Mean age % male Race NR Number Screened, Eligible, Enrolled, Esophagitis Grade (Grading Withdrawn, Lost to Criteria), Other Characteristics Followup Healing Rate at 4 Weeks Healing Rate at 8 Weeks Grade A: 54.2% Screened, eligible esomeprazole 40 mg: NR PP patients (n=42) Grade B: 29.2% NR/48 enrolled omeprazole 20 mg: NR esomeprazole 40 mg: 72.7% Grade C: 8.3% 2 withdrawn/2 lost to omeprazole 20 mg: 50% Grade D: 8.3% followup/42 analyzed (Los Angeles classification) per protocol, 47 ITT patients (n=47) analyzed ITT esomeprazole 40 mg: 64% omeprazole: 20 mg: 45.5% OR (PP: 95% CI , P=0.2040) Proton Pump Inhibitors Page 9 of 168

11 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Chen et al, 2005 Symptoms at 4 Weeks Symptoms at 8 Weeks Results by Baseline Severity NR Heartburn: Not quantitatively expressed, see Figure 1. esomeprazole 40 mg: 50% improved, 50% no change Difference stated as not SS different. omeprazole 20 mg: 65% improved, 25% no change, 10% worse (p=0.0993) Regurgitation: esomeprazole 40 mg: 77.3% improved, 18.2% no change, 4.5% worse omeprazole 20 mg: 85.0% improved, 15.0% no change (p=1.0000) Dysphagia: esomeprazole 40 mg: 36.4% improved, 63.6% no change omeprazole 20 mg: 35.0% improved, 60.0% no change, 5.0% worse (p=0.8697) Epigastric pain: esomeprazole 40 mg: 27.3% improved, 63.6% no change, 9.1% worse omeprazole 20 mg: 50.0% improved, 50.0% no change (p=0.1895) Nausea: esomeprazole 40 mg: 22.7% improved, 68.2% no change, 9.1% worse omeprazole 20 mg: 35.0% improved, 65.0% no change (p=0.5036) Vomiting: esomeprazole 40 mg: 22.7% improved, 77.3% no change omeprazole 20 mg: 40.0% improved, 60.0% no change (p=0.3200) Belching: esomeprazole 40 mg: 54.5%, 36.4% no change, 9.1% worse omeprazole 20 mg: 45.0% improved, 45.0% no change, 10.0% worse (p=0.8999) Withdrawals Due to Adverse Events NR Proton Pump Inhibitors Page 10 of 168

12 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Chen et al, 2005 Quality rating Fair Funding source NR (AstraZeneca provided randomization schedule) Proton Pump Inhibitors Page 11 of 168

13 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Population, Setting Fennerty, patients at multiple centers in the US, with moderate to severe esophagitis. Mean age 47 66% male 82% white, 5% black, <1% Asian, 13% other Number Screened, Eligible, Enrolled, Esophagitis Grade (Grading Withdrawn, Lost to Criteria), Other Characteristics Followup Healing Rate at 4 Weeks Healing Rate at 8 Weeks Grade C: 79% esomeprazole 40 mg: 55.8% esomeprazole 40 mg: 77.5% Grade D: 21% lansoprazole 30 mg: 47.5% lansoprazole 30 mg: 73.3% (Los Angeles classification) (p<0.005) (p=0.099) 4015 screened/ 1381 eligible/ 1001 enrolled/ 11 withdrew/ 18 lost to followup/ 999 analyzed Proton Pump Inhibitors Page 12 of 168

14 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Fennerty, 2005 Symptoms at 4 Weeks Symptoms at 8 Weeks Results by Baseline Severity Resolution of heartburn: Not reported Grade C esomeprazole 40 mg: 72% Healing at 4 weeks lansoprazole 30 mg: 63.6% esomeprazole 40 mg: 60.3% (p=0.005) lansoprazole 30 mg: 50.6% Resolution of acid regurgitation: (p-value not reported) esomeprazole 40 mg: 79.5% Healing at 8 weeks lansoprazole 30 mg: 76.2% esomeprazole 40 mg: 80.3% (p=0.203) lansoprazole 30 mg: 74.9% Dysphagia: (p-value not reported) esomeprazole 40 mg: 93.1% Grade D lansoprazole 30 mg: 93.8% Healing at 4 weeks (p=0.614) esomeprazole 40 mg: 39.8% Epigastric pain: lansoprazole 30 mg: 34.7% esomeprazole 40 mg: 83.1% (p-value not reported) lansoprazole 30 mg: 82.6% Healing at 8 weeks (p=0.831) esomeprazole 40 mg: 67.6% lansoprazole 30 mg: 66.3% (p-value not reported) Withdrawals Due to Adverse Events 5/499 (1%) esomeprazole vs 9/502 (2%) lansoprazole. Most common adverse event leading to study withdrawal was abdominal pain (two in each group) Proton Pump Inhibitors Page 13 of 168

15 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Fennerty, 2005 Quality rating Good Funding source AstraZeneca Proton Pump Inhibitors Page 14 of 168

16 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Population, Setting Gillessen, patients at 27 centers in Germany. Mean age 53 (SD 15) in pantoprazole group, 54 (SD 14) in esomeprazole group. 57% of pantoprazole, 50% of esomeprazole group male. 97% of pantoprazole, 98% of esomeprazole group Caucasian (others Asian) Number Screened, Eligible, Enrolled, Esophagitis Grade (Grading Withdrawn, Lost to Criteria), Other Characteristics Followup Healing Rate at 4 Weeks Healing Rate at 8 Weeks Grade B: 84% pantoprazole, Screened NR/eligible 83% esomeprazole NR/227 enrolled/227 Grade C: 16% pantoprazole, analyzed ITT/197 17% esomeprazole analyzed per protocol (Los Angeles classification) "Early time points" (4 and 6 weeks) Intention-to-treat (N=227): pantoprazole 40 mg: 74% esomeprazole 40 mg: 72% (NS) Per-protocol (N=197): pantoprazole 40 mg: 78% esomeprazole 40 mg: 74% (NS) "Late time points" (8 and 10 weeks) Intention-to-treat (N=227): pantoprazole 40 mg: 90% esomeprazole 40 mg: 92% (NS) Per-protocol (N=197): pantoprazole 40 mg: 96% esomeprazole 40 mg: 93% (NS) Proton Pump Inhibitors Page 15 of 168

17 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Gillessen, 2004 Overall relief of symptoms Per-protocol (N=197): pantoprazole 40 mg: 37% esomeprazole 40 mg: 35% (NS for PP or ITT) Symptoms at 4 Weeks Symptoms at 8 Weeks Results by Baseline Severity Overall relief of symptoms Per-protocol (N=197): pantoprazole 40 mg: 47% esomeprazole 40 mg: 32% (NS for PP or ITT) After 10 weeks: pantoprazole 40 mg: 65% esomeprazole 40 mg: 63% (NS for PP or ITT) Per-protocol, overall healing by baseline grade Grade B: pantoprazole 40 mg: 92% esomeprazole 40 mg: 95% Grade C: pantoprazole 40 mg: 67% esomeprazole 40 mg: 45% Among patients diagnosed with grade C at baseline, 100% of pantoprazole and 91% of esomeprazole improved to Grade A or B at final visit. Withdrawals Due to Adverse Events 6 patients overall, not reported by group. Proton Pump Inhibitors Page 16 of 168

18 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Gillessen, 2004 Quality rating Fair: Randomization, allocation concealment method not reported. Funding source Altana Pharma, Germany Proton Pump Inhibitors Page 17 of 168

19 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Population, Setting Kao et al, patients at one center in Taiwan mean age 49 69% male 100% Asian Number Screened, Eligible, Enrolled, Esophagitis Grade (Grading Withdrawn, Lost to Criteria), Other Characteristics Followup Healing Rate at 4 Weeks Healing Rate at 8 Weeks Grade A: 51% Screened NR/eligible Not reported Not reported Grade B: 49% NR/100 enrolled (Los Angeles Classification) Proton Pump Inhibitors Page 18 of 168

20 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Kao et al, 2003 Symptoms at 4 Weeks Symptoms at 8 Weeks Results by Baseline Severity Esomeprazole 40 mg vs Efficacy of on-demand therapy (n=34 esomeprazole 40 mg, n=23 Not reported omeprazole 20 mg omeprazole 20 mg, initiated week 5) Per-protocol (N=91) Symptom-free on day 1: 28.2% vs 26.2% (NS) Symptom-free before week 1: 56.4% vs 55.6% (NS) Median days to symptom resolution: 4 vs 4 (NS) Achievement of sustained symptom response Week 1: 15.2% vs 15.6% (NS) Week 2: 50% vs 20% (p<0.05) Week 3: 71.7% vs 40% (p<0.01) Week 4: 73.9% vs 51.1% (p<0.05) Week 4 (intention-to-treat): 68% vs 46% (p<0.05) Withdrawals Due to Adverse Events Not reported Proton Pump Inhibitors Page 19 of 168

21 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Kao et al, 2003 Quality rating Fair: not clear if patients masked, randomization, allocation concealment methods not reported. Funding source Supported by a grant from the National Cheng Kung University Proton Pump Inhibitors Page 20 of 168

22 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Castell 1996 Population, Setting 1070 US patients at multiple centers (number excludes placebo), mean age 47, (range 18-84); % male; 85% white, 9% black, 5% Hispanic. Number Screened, Eligible, Enrolled, Esophagitis Grade (Grading Withdrawn, Lost to Criteria), Other Characteristics Followup Healing Rate at 4 Weeks Healing Rate at 8 Weeks Grade 2: 61%-71% 1284 enrolled, 1226 Grade 3: 24%-30% analyzed (total with Grade 4: 6%-9% placebo) (See Appendix F for scale) 6.5%-8.7% Barrett's esophagus lansoprazole 15 mg: 72.0% lansoprazole 30 mg: 79.6% omeprazole 20 mg: 87.0% lansoprazole 30 mg vs lansoprazole 15 mg p<.05 omeprazole 20 mg vs lansoprazole 15 mg p<.05 Other comparisons NS lansoprazole 15 mg: 75.2% lansoprazole 30 mg: 87.1% omeprazole 20 mg: 87.0% lansoprazole 30 mg vs lansoprazole 15 mg p<.05 omeprazole 20 mg vs lansoprazole 15 mg p<.05 Other comparisons NS Castell et al, patients, multiple centers, mean age 47 (range 18-75), 57% male, 91% white, 6% black, 3% other. Grade A: 36% Grade B: 40% Grade C: 18% Grade D: 6% (LA Grade) Heartburn Severity None: 1% Mild: 10% Moderate: 47% Severe: 42% 5241 enrolled, ITT Number screened NR lansoprazole 30 mg (n=2617) esomeprazole 40 mg (n=2624) esomeprazole 79.4% lansoprazole 75.1% (p<.001) (life-table analysis) EE esomeprazole 92.6% lansoprazole 88.8% (p=.0001) (life-table analysis) Proton Pump Inhibitors Page 21 of 168

23 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Castell 1996 Symptoms at 4 Weeks Symptoms at 8 Weeks Results by Baseline Severity Not given Median percentage of days with heartburn: lansoprazole 15 mg: 12.3% lansoprazole 30 mg: 8.6% omeprazole 20 mg: 11.8% Median percentage with heartburn: lansoprazole 15 mg: 9.3 lansoprazole 30 mg: 6.5 (not ITT) lansoprazole15 mg vs omeprazole 20 mg p<0.05 nights lansoprazole15 mg vs lansoprazole 30 mg p< days and nights All other comparisons NS When healing rates were adjusted for baseline esophagitis grade, treatment comparison results were similar to those of the overall analyses. Patients with less severe esophagitis (grade 2) at baseline had higher rates with all the active treatments than those with more severe disease (grades 3 and 4). Healing rate at 4 weeks, lansoprazole 15 mg vs lansoprazole 30 mg vs omeprazole 20 mg, by baseline esophagitis grade: grade 2: 83.2% vs 89.4% vs 88.2% grades 3 and 4: 59.5% vs 73.5% vs 69.8% at 8 weeks, lansoprazole 15 mg vs lansoprazole 30 mg vs omeprazole 20 mg, by baseline esophagitis grade:: grade 2: 87.8% vs 94.3% vs 91.6% grades 3 and 4: 62.5% vs 85.3% vs 88.7% Withdrawals Due to Adverse Events omeprazole 20 mg: 2% lansoprazole 30 mg: 1.7% lansoprazole 15 mg: 0.9% Castell et al, 2002 Complete resolution of heartburn: lansoprazole 60.2% esomeprazole 62.9% (p<.05) Not reported esomeprazole 75.7% lansoprazole 71.7% (p<0.01, stratified by baseline severity) No difference in treatment-related adverse effects. Heartburn-free nights: lansoprazole 85.8% esomeprazole 87.1% (p<.05) esomeprazole 87.6% lansoprazole 84.2% (p<0.01, stratified by baseline severity) Withdrawal due to adverse event 1.8% vs. 1.9%. Heartburn-free days: NS Proton Pump Inhibitors Page 22 of 168

24 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Castell 1996 Quality rating Fair: randomization and allocation method not reported, attrition not reported Funding source Supported by TAP Pharmaceuticals, Inc. Castell et al, 2002 Good Supported by AstraZeneca, also listed in author credits Proton Pump Inhibitors Page 23 of 168

25 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Corinaldesi 1995 Population, Setting 241 patients at 30 centers, Belgium, France, Italy, the Netherlands, median age 50-52, (range 18-88); 63% male; ethnicity not given. Number Screened, Eligible, Enrolled, Esophagitis Grade (Grading Withdrawn, Lost to Criteria), Other Characteristics Followup Healing Rate at 4 Weeks Healing Rate at 8 Weeks Grade 2: 82% Number screened not pantoprazole 40 mg: 67.5% pantoprazole 40 mg: 80.8% Grade 3: 18% given, 241 omeprazole 20 mg: 68.6% omeprazole 20 mg: 79.3% (Savary-Miller) randomized, 208 p=ns p=ns evaluable; 3 withdrew, 23 did not attend f/u. Proton Pump Inhibitors Page 24 of 168

26 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Corinaldesi 1995 Symptoms at 4 Weeks Symptoms at 8 Weeks Results by Baseline Severity Withdrawals Due to Adverse Events Heartburn free: omeprazole 20 mg: 82.2% pantoprazole 40 mg: 87.9% p=ns Not reported Not reported pantoprazole 40 mg: 0.8% omeprazole 20 mg: 1.7% Proton Pump Inhibitors Page 25 of 168

27 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Corinaldesi 1995 Quality rating Funding source Poor: randomization and allocation method not Last author from reported, no intention-to-treat analysis, baseline Byk Gulden Pharmaceuticals, study characteristics not analyzed. supported by same. Proton Pump Inhibitors Page 26 of 168

28 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Dekkers 1999 Number Screened, Eligible, Enrolled, Esophagitis Grade (Grading Withdrawn, Lost to Population, Setting Criteria), Other Characteristics Followup Healing Rate at 4 Weeks Healing Rate at 8 Weeks 202 patients of 27 Grade 2: 43% Number screened not rabeprazole 20 mg: 81% rabeprazole 20 mg: 92% investigators in 10 Grade 3: 52% given, 202 enrolled, omeprazole 20 mg: 81% omeprazole 20 mg: 94% European countries, mean Grade 4: 4% 192 completed. (Not ITT) (Not ITT) age , (range 20- (modified Hetzel-Dent) p=ns p=ns 86); 62% male; ethnicity not given. Delchier patients of 61 investigators at 50 European centers, mean age 53 (+15), (range 18-80); 62% male; ethnicity not given. Mean grade , median 3.9, (modified Hetzel-Dent) 7% had Barrett's esophagus, 41% positive for H. pylori 358 screened, 310 randomized, 298 completed. rabeprazole 20 mg: 88.5% rabeprazole 10 mg: 85.4% omeprazole 20 mg: 91.2% p=ns rabeprazole 20 mg: 91.3% rabeprazole 10 mg: 91.3% omeprazole 20 mg: 94.2% p=ns Proton Pump Inhibitors Page 27 of 168

29 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Dekkers 1999 Symptoms at 4 Weeks Symptoms at 8 Weeks Results by Baseline Severity Heartburn frequency (resolution): Heartburn frequency resolution: Not reported rabeprazole 20 mg: 29.6% rabeprazole 20 mg: 37.8% omeprazole 20 mg: 26.5% omeprazole 20 mg: 31.4% Daytime severity (resolution): Daytime severity resolution: rabeprazole 20 mg: 61.9% rabeprazole 20 mg:68.0% omeprazole 20 mg: 60.8% omeprazole 20 mg: 66.0% Nighttime severity resolution: Nighttime severity resolution: rabeprazole 20 mg: 61.6% rabeprazole 20 mg: 64.4% omeprazole 20 mg: 57.3% omeprazole 20 mg: 66.7% p=ns for all p= NS for all Withdrawals Due to Adverse Events rabeprazole 20 mg: 1% omeprazole 20 mg: 0 Delchier 2000 Severity of daytime and nighttime heartburn: p=ns (numbers not given) Severity of daytime and nighttime heartburn: p=ns (numbers not given) No statistically significant differences between rabeprazole 10 mg: treatment groups after controlling for baseline 5% factors including Hetzel-Dent grade (other factors rabeprazole 20 mg: sex, age, smoking and H. pylori status); data not 5% reported. omeprazole 20 mg: 2% Proton Pump Inhibitors Page 28 of 168

30 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Dekkers 1999 Quality rating Fair: randomization and allocation method not reported intention-to-treat for symptoms only, not for healing. Funding source Last author (corresponding author) and 5th authors with Eisai Ltd, funding info not given. Delchier 2000 Fair: randomization and allocation method not reported, followup somewhat high (76%-83%). Funded by Eisai Ltd, London, last author (corresponding author) from Eisai Proton Pump Inhibitors Page 29 of 168

31 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Dupas 2001 Population, Setting 461 patients at 29 hospital centers and 45 private practices in France; mean age 54 (+14.6); 74% male; ethnicity not given Number Screened, Eligible, Enrolled, Esophagitis Grade (Grading Withdrawn, Lost to Criteria), Other Characteristics Followup Healing Rate at 4 Weeks Healing Rate at 8 Weeks 83% Grade 2 17% Grade 3 (Savary-Miller) Number screened not given; 461 randomized, 385 completed pantoprazole 40 mg ITT: 80.90% lansoprazole 30 mg ITT: 80% p=ns pantoprazole 40 mg ITT: 89.80% lansoprazole 30 mg ITT: 90% p=ns Proton Pump Inhibitors Page 30 of 168

32 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Dupas 2001 Symptoms at 4 Weeks Symptoms at 8 Weeks Results by Baseline Severity Symptom free (all symptoms - Not reported heartburn, acid regurgitation, pain or swallowing): ITT: pantoprazole 40 mg: 83% lansoprazole 30 mg: 92% p=ns For both treatments, healing rates after 4 weeks were lower in grade III than in grade II esophagitis (69% vs 89%, per-protocol analysis, p=0.0001), with no grade-dependent significant differences between groups. Withdrawals Due to Adverse Events pantoprazole 40 mg: 13% lansoprazole 30 mg: 2.5% Proton Pump Inhibitors Page 31 of 168

33 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Dupas 2001 Quality rating Fair: randomized method not clear, allocation method not reported Funding source Funded by BYK France, last author from BYK Proton Pump Inhibitors Page 32 of 168

34 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Hatlebakk 1993 Population, Setting 229 patients at 9 hospitals in Norway and Sweden; mean age 55; 66% male; ethnicity not given Number Screened, Eligible, Enrolled, Esophagitis Grade (Grading Withdrawn, Lost to Criteria), Other Characteristics Followup Healing Rate at 4 Weeks Healing Rate at 8 Weeks lansoprazole 30 mg group: Number screened not lansoprazole 30 mg: 61.2% lansoprazole 30 mg: 81.9% Grade 0: 2.6% given, 229 enrolled. omeprazole 20 mg: 64.6% omeprazole 20 mg: 85.0% Grade 1: 34.5% p=ns p=ns Grade 2: 50.9% Grade 3: 12.1% omeprazole 20 mg group: Grade 0: 2.7% Grade 1: 38.9% Grade 2: 55.8% Grade 3: 2.7% (See Appendix E for scale) Holtmann, patients at multiple centers in Germany, Denmark, and Switzerland; mean age 52; 66% male, 99% Caucasian. rabeprazole: 78% grade II, 22% grade III; omeprazole: 84% grade II, 16% grade III 274 screened/254 eligible, 251 enrolled/13 withdrawn or no valid data/4 lost to followup/251 analyzed No difference between groups (data not reported) per protocol (N=200) rabeprazole 20 mg: 92.7% omeprazole 40 mg: 89.2% (NS) Proton Pump Inhibitors Page 33 of 168

35 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Hatlebakk 1993 Symptoms at 4 Weeks Symptoms at 8 Weeks Results by Baseline Severity Data not given: Data not given, but states no significant differences in any states lansoprazole 30 mg had symptoms. greater improvement in heartburn (p=0.03) At both 4 and 8 weeks, and irrespective of treatment, healing rates were higher for patients with grade 1 esophagitis than grade 2 (p<0.01, two-stage logistic regression analysis). Results by treatment group not reported. Withdrawals Due to Adverse Events omeprazole 20 mg: 0.9% lansoprazole 30 mg: 0 Holtmann, 2002 Not reported for this time point; difference in relief from heartburn on day 4 not significant between groups. Not reported for this time point. Healing rate in patients with GERD grade III (N=45) 4 weeks: 84% rabeprazole vs 72.2% omeprazole (NS) 8 weeks: 88% rabeprazole vs 77.8% omeprazole (NS) 4/125 (3%) rabeprazole vs 2/126 (2%) omeprazole Proton Pump Inhibitors Page 34 of 168

36 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Hatlebakk 1993 Quality rating Funding source Poor: randomization and allocation method not Not reported reported, no intention-to-treat analysis, eligibility criteria not specified, some differences at baseline. Holtmann, 2002 Fair: Not clear if randomization method adequate, allocation concealment method not reported, more rabeprazole patients grade III esophagitis at baseline (22% vs 16%). Funded by Eisai and Janssen-Cilag Proton Pump Inhibitors Page 35 of 168

37 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Howden et al, 2002 Population, Setting 284 patients at multiple centers, mean age 46.5 (range 19-78), 56% male, 80% white, 5% black, 15% other. Number Screened, Eligible, Enrolled, Esophagitis Grade (Grading Withdrawn, Lost to Criteria), Other Characteristics Followup Healing Rate at 4 Weeks Healing Rate at 8 Weeks Grade 2: 61% 284 enrolled; # lansoprazole 30 mg vs lansoprazole 30 mg vs Grade 3:30% screened, eligible not esomeprazole 40 mg esomeprazole 40 mg Grade 4: 8% reported, % vs 78.3% (p=ns) 91.4% vs 89.1% (see Appendix F for scale) evaluated (95% CI of difference -4.7, 9.2) lansoprazole 30 mg (n=139) esomeprazole 40 mg (n=138) Proton Pump Inhibitors Page 36 of 168

38 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Howden et al, 2002 Symptoms at 4 Weeks Symptoms at 8 Weeks Results by Baseline Severity Withdrawals Due to Adverse Events Not reported Not reported Healing rate or improvement of 2 grades at 8 2/143 (1.4%) weeks by baseline grade, lansoprazole 30 mg vs lansoprazole vs 5/141 esomeprazole 40 mg: (3.5%) esomeprazole Grade 2: 94.3% (82/87) vs 95.1% (77/81) Grade 3: 92.7% (38/41) vs 81.8% (36/44) Grade 4: 90.9% (10/11) vs 84.6% (11/13) Week 4 healing: healing or improvement of 2 grades of erosive esophagitis from baseline were comparable between treatment groups, regardless of baseline grade of esophagitis (data not reported). Proton Pump Inhibitors Page 37 of 168

39 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Howden et al, 2002 Quality rating Funding source Fair: randomization and allocation concealment Supported by TAP methods not reported. Pharmaceuticals. Proton Pump Inhibitors Page 38 of 168

40 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Kahrilas 2000 Population, Setting 1960 US patients at 140 centers; mean age 46; 60% male; ethnicity not given. Number Screened, Eligible, Enrolled, Esophagitis Grade (Grading Withdrawn, Lost to Criteria), Other Characteristics Followup Healing Rate at 4 Weeks Healing Rate at 8 Weeks Grade A: 33% Grade B: 40% Grade C: 19% Grade D: 7% (Los Angeles classification) 9.6% H. pylori 3354 screened, 1960 randomized. 44 did not complete study due to an adverse event and 115 for other reasons including loss to f/u and withdrawal of consent. esomeprazole 40 mg: 75.9% esomeprazole 20 mg: 70.5% omeprazole20: 64.7% (cumulative life table rate) esomeprazole 20 mg vs omeprazole 20 mg p=0.09 esomeprazole 40 mg vs omeprazole 20 mg (p <0.05) esomeprazole 40 mg: 94.1% esomeprazole 20 mg: 89.9% omeprazole 20 mg: 86.9% (cumulative life table rate) esomeprazole 40 mg vs omeprazole 20 mg p<0.001 esomeprazole 20 mg vs omeprazole 20 mg p<0.05 Proton Pump Inhibitors Page 39 of 168

41 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Kahrilas 2000 Symptoms at 4 Weeks Symptoms at 8 Weeks Results by Baseline Severity Resolution of heartburn "Cumulative analysis at week 8 not done because pts could complete Greater efficacy of esomeprazole 40 mg vs esomeprazole 40 mg: 64.7% the study at week 4 with healed reflux esophagitis, even if symptoms omeprazole 20 mg at 4 weeks was consistent esomeprazole 20 mg: 61.0% were present" when adjusting for baseline esophagitis grade omeprazole 20 mg: 57.2% (data not reported). esomeprazole 40 mg vs omeprazole 20 mg p=0.005 other comparisons NS Withdrawals Due to Adverse Events esomeprazole 40 mg: 2% esomeprazole 20 mg: 2.6% omeprazole 20 mg: 2% Proton Pump Inhibitors Page 40 of 168

42 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Kahrilas 2000 Quality rating Funding source Fair: Randomization methods not reported, 4 of 9 authors from baseline characteristics not analyzed, more Astra Zeneca, study grade A patients (mild) in esomeprazole 40 mg supported by grant group than omeprazole 20 mg group at from Astra Zeneca. baseline (35.9% esomeprazole vs 31.2% omeprazole 20 mg; calculated p = 0.07). Proton Pump Inhibitors Page 41 of 168

43 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Korner et al, 2003 Population, Setting 669 patients at multiple centers, mean age 53.8 (sd 14), 60% male, ethnicity not reported. Number Screened, Eligible, Enrolled, Esophagitis Grade (Grading Withdrawn, Lost to Criteria), Other Characteristics Followup Healing Rate at 4 Weeks Healing Rate at 8 Weeks 84% Grade II 669 included; number 16% Grade III screened, eligible not (Savary-Miller) reported. Pantoprazole 40 mg (n=337) omeprazole MUPS 40 mg (n=332) ITT results reported as odds ratios only. PP results, pantoprazole 40 mg (n=282) vs omeprazole MUPS 40 mg (n=270) 70.9% vs 72.6% ITT results reported as odds ratios only. "Healing rates after 8 weeks of treatment were also similar in both groups." Labenz et al, patients, multinational, mean age 50.6 (sd 14), 63% male, 97% Caucasian. Grade A: 32% Grade B: 44% Grade C: 19% Grade D: 5% (LA Classification) 3170 randomized, 3151 analyzed. 9 excluded from analysis because of intake of an unknown study drug, and 10 because of study protocol violations. esomeprazole 40 mg vs pantoprazole 40 mg Observed (per protocol): 78.8% vs 72.8% risk difference 6% (95% CI 3%, 9%) Life table analysis, per protocol: 81.0% vs 74.5% (p<0.001) esomeprazole 40 mg vs pantoprazole 40 mg Observed (per protocol): 91.6% vs 88.9% risk difference 3% (95% CI 1%, 5%) Life table analysis, per protocol: 95.5% vs 92.0% (p<0.001) Proton Pump Inhibitors Page 42 of 168

44 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Korner et al, 2003 Symptoms at 4 Weeks Symptoms at 8 Weeks Results by Baseline Severity Withdrawals Due to Adverse Events ITT results not reported PP, pantoprazole 40 mg vs omeprazole MUPS 40 mg: Heartburn relief: 83.7% vs 88.1% Relief of pain on swallowing: 83.1% vs 91.9% (p-values not reported) ITT results not reported PP, pantoprazole 40 mg vs omeprazole MUPS 40 mg: Heartburn relief: 91.1% vs 92.6% Relief of pain on swallowing: 94.1% vs 96.3% (p-values not reported) Not reported (all patients were Grade II or III) 4/337 (1%) pantoprazole, 7/332 (2%) omeprazole MUPS Labenz et al, 2005 esomeprazole 40 mg vs pantoprazole 40 mg Time to achieve sustained heartburn resolution (defined as the first of 7 consecutive days with no heartburn): 6 days vs 8 days (p<0.001) esomeprazole 40 mg vs pantoprazole 40 mg Proportion of heartburn-free days: 70.7% vs 67.3% (p<0.01) Healing of esophagitis by baseline grade, esomeprazole 40 mg vs pantoprazole 40 mg Week 4, (Observed, per protocol): Grade A: 83.9% vs 83.1% (NS) Grade B: 80.2% vs 75.4% (p<0.05) Grade C: 71.1% vs 60.1% (p<0.01) Grade D: 61.4% vs 40.2% (p<0.01) 2.1% esomeprazole, 1.8% pantoprazole Week 8 (Life table analysis, per protocol): Grade A: 97.3% vs 97.1% (NS) Grade B: 96.9% vs 93.1% (p<0.05) Grade C: 91.3% vs 87.6% (p<0.01) Grade D: 88.1% vs 73.6% (p<0.05) Proton Pump Inhibitors Page 43 of 168

45 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Korner et al, 2003 Quality rating Fair: ITT results not reported, randomization and allocation concealment methods not reported. Funding source Supported by a grant from ALTANA Pharma AG, Germany. Labenz et al, 2005 Fair/Poor: AstraZeneca Randomization and allocation concealment methods not reported. Post-randomization exclusions (19 patients) and no data on excluded patients. Baseline data excludes 19 patients randomized but excluded due to intake of an unknown study drug or protocol violations. No data on excluded patients. Some differences in baseline esophagitis grade at baseline (grade B: 42.6% esomeprazole vs 45.1% pantoprazole; grade D: 4.5% esomeprazole, 5.8% pantoprazole). Proton Pump Inhibitors Page 44 of 168

46 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Pace et al, 2005 Population, Setting 549 patients, multi center Italy, mean age 47.4 (sd 14), male 68.1% Number Screened, Eligible, Enrolled, Esophagitis Grade (Grading Withdrawn, Lost to Criteria), Other Characteristics Followup Healing Rate at 4 Weeks Healing Rate at 8 Weeks Grade 0: 1% Screened NR, Eligible rabeprazole 20 mg: PP 91.0%, rabeprazole 20 mg: PP 97.9%, Grade 1: 69% NR, Enrolled 560, omeprazole 20 mg: PP 89.9%, omeprazole 20 mg: PP 97.5%, Grade 2: 24% Withdrawn 47, lost to equivalence bet. the two drugs is equivalence bet. the two drugs Grade 3: 5.5% f/u 9 statistically significant (p<0.001) is statistically significant Grade 4: 0% (p<0.0001) (Savary-Miller) Proton Pump Inhibitors Page 45 of 168

47 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Pace et al, 2005 Symptoms at 4 Weeks Symptoms at 8 Weeks Results by Baseline Severity ITT population, mean time to the NR first day w/ satisfactory heartburn relief, rabeprazole (n=271) days, omeprazole (n=271) days (p=0.0045), mean time to complete heartburn relief, rabeprazole 7.2 days, omeprazole 8.4 days (p=ns). Patients w/ complete heartburn relief (day and nighttime) in each day of first week of treatment (ITT patients) Rabeprazole n= %, Omeprazole n= % Healing rates of oesophagitis grade at endpoint (4 or 8 weeks), rabeprazole vs omeprazole: grade I: 99.4 vs. 98.8%, grade II: 95.1 vs. 96.4%, grade III: 91.7 vs. 86.7% (PP patients) Withdrawals Due to Adverse Events No significant difference bet. Treatment groups in single adverse event occurring, with exception of headache (Omeprazole 4.8% and Rabeprazole 1.4%) Proton Pump Inhibitors Page 46 of 168

48 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Pace et al, 2005 Quality rating Fair. Lack of ITT analysis, exclusion of people (2%) at baseline. Funding source Janssen-Cilag, Italy Proton Pump Inhibitors Page 47 of 168

49 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Mee 1996 Population, Setting 604 patients at multiple centers, UK and Ireland, mean age 53; 67% male; ethnicity not given. Number Screened, Eligible, Enrolled, Esophagitis Grade (Grading Withdrawn, Lost to Criteria), Other Characteristics Followup Healing Rate at 4 Weeks Healing Rate at 8 Weeks Grade 1: 39% 604 enrolled, 565 lansoprazole 30 mg: 62% lansoprazole 30 mg: 75.3% Grade 2: 44% eligible, 537 evaluable omeprazole 20 mg: 56.6% omeprazole 20 mg: 71.1% Grade 3: 15% p=ns p=ns Grade 4: 2% (Savary-Miller) Proton Pump Inhibitors Page 48 of 168

50 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Mee 1996 Symptoms at 4 Weeks Symptoms at 8 Weeks Results by Baseline Severity Not given Improvement in daytime epigastric pain Healing of esophagitis by baseline grade, lansoprazole 30 mg: 85.9% lansoprazole vs omeprazole: omeprazole 20 mg: 72.5% Week 4: Improvement in nighttime epigastric pain Grade I: 79% vs 68% lansoprazole 30 mg: 85.9% Grade II: 72% vs 62% omeprazole 20 mg: 67.3% Grade III: 45% vs 57% p=ns Grade IV: 43% vs 60% (includes only pts who attended 8-week visit who reported baseline Week 8 (cumulative): pain) Grade I: 92% vs 87% Grade II: 88% vs 81% Grade III: 73% vs 72% Grade IV: 50% vs 50% Esophagitis grade and treatment were included in a logistic regression model. Odds ratio of healing on lansoprazole compared with omeprazole was 1.46 (95% CI 0.87, 2.45) Withdrawals Due to Adverse Events Not reported Proton Pump Inhibitors Page 49 of 168

51 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Mee 1996 Quality rating Good/Fair: Allocation concealment method not given. Funding source 1 of 2 authors from Lederle Laboratories, funding info not given. Proton Pump Inhibitors Page 50 of 168

52 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Mulder 1996 Population, Setting 211 patients at multiple centers in The Netherlands; mean age 55; 70% male; ethnicity not given. Number Screened, Eligible, Enrolled, Esophagitis Grade (Grading Withdrawn, Lost to Criteria), Other Characteristics Followup Healing Rate at 4 Weeks Healing Rate at 8 Weeks Grade 1: 0.47% (1 patient) Grade 2: 68% Grade 3: 24% Grade 4A: 8% (Savary-Miller) Number screened not given, 211 enrolled, 3 lost to followup, 3 withdrew for lack of efficacy, 1 withdrawn for receiving double dose. lansoprazole 30 mg ITT 85.50% PP 86.20% omeprazole 40 mg ITT 79% PP 79.6% p=ns lansoprazole 30 mg ITT: 93.40% PP 95.70% omeprazole 40 mg ITT: 90.50% PP 93.4% p=ns Proton Pump Inhibitors Page 51 of 168

53 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Mulder 1996 Symptoms at 4 Weeks Symptoms at 8 Weeks Results by Baseline Severity lansoprazole 30 mg "Because of the low number of patients not healed at 4 weeks, Healing of esophagitis by baseline grade, No symptoms: analysis of symptoms was not performed at 8 weeks." lansoprazole vs omeprazole: ITT: Week 4: 73.60% Grade II: 90.8% vs 88.1% omeprazole 40 mg Grade III/IV: 81.5% vs 70.6% No symptoms: overall: ITT Grade II: 97.4% vs 98.5% 71.40% Grade III/IV: 92.6% vs 85.3% (All NS) Withdrawals Due to Adverse Events None Proton Pump Inhibitors Page 52 of 168

54 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Mulder 1996 Quality rating Funding source Fair: randomization and allocation concealment Supported by not reported, Hoechst Marion Roussel BV and Janssen-Cilag BV, Netherlands Proton Pump Inhibitors Page 53 of 168

55 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Mulder et al Population, Setting 461 patients, multiple centers; mean age 51.2 (range 18-80);59% male; ethnicity NR Number Screened, Eligible, Enrolled, Esophagitis Grade (Grading Withdrawn, Lost to Criteria), Other Characteristics Followup Healing Rate at 4 Weeks Healing Rate at 8 Weeks Savary-Miller class: 461 enrolled NR NR I: 59% II: 29% Number screened NR III: 8% IVa: 4% Heartburn Severity None: 4% Mild: 22% Moderate: 45% Severe: 29% omeprazole MUPS 20 mg (n=151) lansoprazole 30 mg (n=156) pantoprazole 40 mg (n=154) Proton Pump Inhibitors Page 54 of 168

56 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Mulder et al Symptoms at 4 Weeks Symptoms at 8 Weeks Results by Baseline Severity (omeprazole vs lansoprazole vs (omeprazole vs lansoprazole vs pantoprazole) pantoprazole) Heartburn relief : 87% vs. 81% vs. 89% Heartburn relief : 84% vs. 78% vs. pantoprazole vs omeprazole 90% CI to % omeprazole vs lansoprazole 90% CI to omeprazole vs lansoprazole 90% CI pantoprazole vs lansoprazole 90% CI 0.94 to to Satisfied: 89% vs. 86% vs. 91% pantoprazole vs lansoprazole 90% omeprazole vs lansoprazole 90% CI to 9.69 CI to pantoprazole vs lansoprazole 90% CI to Satisfied: 79% vs. 76% vs. 79%. pantoprazole vs omeprazole 90% CI to 7.13 omeprazole vs lansoprazole 90% CI to pantoprazole vs lansoprazole 90% CI to pantoprazole vs omeprazole 90% ci to 7.13 Symptom relief at 4 and 8 weeks was similar for each grade of esophagitis. Maintenance phase (with omeprazole 20 mg or 40 mg only, N=391): symptom relief with omeprazole 20 mg was independent of initial severity of esophagitis; the number of patients in the omeprazole 40 mg maintenance group (N=21) was too small to be divided by initial esophagitis grade. Withdrawals Due to Adverse Events No difference in AEs between groups. None considered treatment related. Total withdrawals due to AE: 6/461 (1.3%) Total AEs: 73/461 (15.8%) Proton Pump Inhibitors Page 55 of 168

57 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Mulder et al Quality rating Fair: randomization and allocation methods not reported. More withdrawals in L group. Funding source Supported by AstraZeneca Proton Pump Inhibitors Page 56 of 168

58 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Richter et al, 2001a Population, Setting 2425 patients at 163 US centers; mean age 47 (sd 12); 61% male; ethnicity 93.5% Caucasian. Number Screened, Eligible, Enrolled, Esophagitis Grade (Grading Withdrawn, Lost to Criteria), Other Characteristics Followup Healing Rate at 4 Weeks Healing Rate at 8 Weeks Grade A: esomeprazole 40 mg 4798 screened, 2425 esomeprazole 40 mg vs esomeprazole 40 mg vs 35%; omeprazole 20 mg 32% randomized; 109 did omeprazole 20 mg omeprazole 20 mg Grade B: esomeprazole 40 mg not complete: 24 for cumulative life table rate: cumulative life table rate: 39%; omeprazole 20 mg 42% adverse events, % vs 68.7% (p<0.001) 93.7% vs 84.2% (p<0.001) Grade C: esomeprazole 40 mg investigator-initiated 21%; omeprazole 20 mg 20% decision, 25 lost to Crude rates: Crude rates: Grade D: esomeprazole 40 mg followup, 31 consent 78.6% vs 66.6% (p = for 89.9% vs 81.0% (p = for 5%; omeprazole 20 mg 7% withdrawn, 4 lack of CMH test) CMH test) (LA classification) therapeutic response. risk difference 12% (95% CI 9%, risk difference 9% (95% CI 6%, 16%) 12%) Proton Pump Inhibitors Page 57 of 168

59 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Richter et al, 2001a Symptoms at 4 Weeks Symptoms at 8 Weeks Results by Baseline Severity esomeprazole 40 mg "Cumulative analysis at week 8 not done because pts could complete resolution of heartburn: the study at week 4 with healed reflux esophagitis, even if symptoms 68.30% were present" omeprazole 20 mg resolution of heartburn: 58.10% Greater efficacy of esomeprazole 40 mg vs omeprazole 20 mg at 4 weeks was consistent when adjusting for baseline esophagitis grade. Week 4 healing rates by baseline esophagitis grade (approximate, estimated from figure): esomeprazole 40 mg vs omeprazole 20 mg: Grade A: 88% vs 82% Grade B: 79% vs 66% Grade C: 71% vs 53% Grade D: 55% vs 35% Week 8 healing rates by baseline esophagitis grade (approximate, estimated from figure): esomeprazole 40 mg vs omeprazole 20 mg: Grade A: 93% vs 91% Grade B: 90% vs 82% Grade C: 88% vs 70% Grade D: 80% vs 62% (p=0.001 for CMH test, esomeprazole vs omeprazole) Withdrawals Due to Adverse Events 1% in each group Proton Pump Inhibitors Page 58 of 168

60 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Richter et al, 2001a Quality rating Good Funding source Supported by Astra Zeneca, one or more authors from Astra Zeneca. Proton Pump Inhibitors Page 59 of 168

61 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Richter et al., 2001b Population, Setting 3510 patients, multiple centers, mean age 47 (range 18-89); 57% male, 88% white, 5% black, 7% other. Number Screened, Eligible, Enrolled, Esophagitis Grade (Grading Withdrawn, Lost to Criteria), Other Characteristics Followup Healing Rate at 4 Weeks Healing Rate at 8 Weeks Grade 0: <1% 3410 enrolled; number Not evaluated Not evaluated Grade 1: 0% screened, eligible not Grade 2: 68% reported. Grade 3: 25% Grade 4: 7% (See Appendix F for scale) Scholten et al., patients at multiple centers, mean age 53 (sd ~14); 99% white Grade B: 73% Grade C: 27% (LA Classification) 217 enrolled; number screened, eligible not reported. Not evaluated Not evaluated Proton Pump Inhibitors Page 60 of 168

62 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Richter et al., 2001b Symptoms at 4 Weeks Symptoms at 8 Weeks Results by Baseline Severity Withdrawals Due to Adverse Events lansoprazole 30 mg vs omeprazole 20 mg Sustained resolution of heartburn: 77.2% vs 76.2% (p=ns) lansoprazole 30 mg vs omeprazole 20 mg Sustained resolution of heartburn: 84.3% vs 83.0% (p=ns) More patients talking lansoprazole did not have a single episode of day or night heartburn (between 10% and 15%, p<0.05, data are presented graphically only) Not reported 40/1754 (2%) lansoprazole 33/1756 (2%) omeprazole. Scholten et al., 2003 pantoprazole 40 mg vs esomeprazole 40 mg No or only mild heartburn: 99% vs 98% Not evaluated Not reported (all patients were Grade B or C) 3 patients discontinued due to adverse events not related to study drug (myocardial infarction, headache, allergic reaction). Groups not reported. Proton Pump Inhibitors Page 61 of 168

63 Evidence Table 1. Erosive GERD short-term trials of PPI vs PPI Richter et al., 2001b Quality rating Fair: ITT results not reported, randomization and allocation concealment methods not reported. Funding source Supported by a grant from TAP Pharmaceuticals Scholten et al., 2003 Fair: ITT results not reported, randomization and allocation concealment methods not reported. Supported by a grant from ALTANA Pharma AG, Germany. Proton Pump Inhibitors Page 62 of 168

64 Evidence Table 2. Quality assessment of included trials Internal Validity, Country Randomization adequate? Allocation concealment adequate? Groups similar at baseline? Eligibility criteria specified? Outcome assessors masked? Care provider masked? Patient masked? Adachi 2003 Method not reported Yes Yes Yes No- open No No Ando 2005 Method not reported Not reported Some Yes Yes Yes Yes Armstrong et al., 2004 Method not reported Not reported Yes Yes Described as double-blind, not specified Described as double-blind, not specified Described as double-blind, not specified Bardhan 2001 Method not reported Not reported More smokers in Yes No- open No No pantoprazole group (31% vs 22%), more males in omeprazole group (64% vs 52%) Bytzer et al., 2004 Method not reported Not reported Yes Yes Yes Yes Yes Proton Pump Inhibitors Page 63 of 168

65 Evidence Table 2. Quality assessment of included trials, Country Adachi 2003 Reporting of attrition, crossovers, adherence, and contamination Attrition and adherence yes Loss to follow-up: differential/high Yes- 20% of lansoprazole group lost to f/u for endoscopy 7% in other groups; but no loss to f/u for reporting of symptoms. Intention-to-treat (ITT) analysis Postrandomization exclusions Quality Rating Yes for symptoms No Fair-poor Ando 2005 Armstrong et al., 2004 attrition yes, adherence no, crossovers no, contamination no No Not reported Unable to determine (defined as all randomized patients who took at least one dose of study medication and had postrandomization data, but number withdrawn not reported) No No Yes Fair Unable to determine Fair Bardhan 2001 Attrition and adherence yes No Yes No Fair Bytzer et al., 2004 Attrition yes, others no No No (analyzed patients who had data on at least 1 postrandomization visit; number not specified) No Fair Proton Pump Inhibitors Page 64 of 168

66 Evidence Table 2. Quality assessment of included trials Internal Validity, Country Caos et al., 2005 Randomization adequate? Allocation concealment adequate? Groups similar at baseline? Eligibility criteria specified? Outcome assessors masked? Care provider masked? Patient masked? Yes Not reported Yes Yes Yes Yes Yes Chen, 2005 Yes Not reported omeprazole group older (59.0 vs 49.2, p=0.0596), more belching in esomeprazole group (47% vs 25.2%, p=0.0121) Yes Yes Described as double-blind, not specified Yes (placebo) Cucchiara 1993 Method not reported Not reported Few given, some Yes Some No No differences - clinical significance unclear Fennerty 2005 Yes Yes Yes Yes Yes Yes Yes Florent 1994 Method not reported Not reported More patients with previous hemorrhage in O group Yes Unclear Unclear Unclear Fock et al., 2005 Yes Method not reported More women in Yes esomeprazole group (57.8% vs 39.7%, p=0.051); otherwise similar Described as double-blind, tablets inserted in identical capsules Described as double-blind, tablets inserted in identical capsules Described as double-blind, tablets inserted in identical capsules Gillessen 2004 Method not reported Not reported Yes Yes Yes Yes Yes Holtmann 2001 Not clear if adequate method Not reported 22% of rabeprazole group Grade III vs 16.4% omeprazole Yes Yes Yes Yes Kao 2003 Method not reported Not reported Yes Yes Yes Not reported Not clear Proton Pump Inhibitors Page 65 of 168